i am using bryan chungs face detection library and example to build a sketch that switches all faces detected out with eachother.

im having trouble with the flickering of the face detection and want the faces to fade out instead of on/off approach
i also want to mask out the faces with a masking image this is done easily with img.mask.pixels kinda way.

anyway here is some code:
Code:

void drawFace() {
  int k;
  int [][] res = face.getFaces();
    for (int i=0;i<res.length;i++) {
	int x = res[i][0];
	int y = res[i][1];
	int w = res[i][2];
	int h = res[i][3];

	images[i] = get(x,y,w,h);
	
	/* here the mask is resized and made into an array of pixels so 
	it can be used by multiple faces at once*/
	mymask2=createImage(w,h,RGB);
	mymask2.copy(mymask,0,0, 300, 300,x,y,w,h);
		mymask2.loadPixels();
	mymask2.updatePixels();

     k=i;
/*here three different states is calculated, one face, move first face to last face, and loop down the faces*/
     if(res.length==1){
	 images[i].mask(mymask2.pixels);
	 image(images[i],x,y,w,h);
	 println("one face");
     }else{
	 	  
	 if(i-1==-1){
	     images[i].mask(mymask2.pixels);
	   image(images[i],res[0][0],res[0][1],res[0][2],res[0][3]);
	   println("more faces 2");

	 }else{
	     images[i].mask(mymask2.pixels);
	  image(images[i],res[i-1][0],res[i-1][1],res[i-1][2],res[i-1][3]);
   println("more faces 1");
	 }  
    }
  }
  } 

if there is any help at all you can give me dont hesitate

I've got very limited experience with the face-detection libraries; though certainly found examples to mask out the background from one face (though things obviously get more complex when working with more than one).

Face recognition (as opposed to detection) is non-trivial; but I think what you really mean is slightly different.  IIRC face detection libraries don't necessarily return the array of recognised faces in the same order each frame, meaning that a face might jump from index 0 to index 3 and back again in a matter of a few frames - which obviously causes problems with your goal.

To resolve the problem I guess I'd create a 'Face' class with x and y position.  When a new face is recognised I'd create a Face object.  Then each frame I'd compare the position of faces returned by the face detection library with last known position of the Face objects and work on the crude basis that the nearest detected face to the last known position of the Face object is the correct one.  You'd then do the face swapping on the Face object rather than the detected face.

That's a pretty crude solution and is open to all kinds of problems: like what happens when two faces get close together?  How do you determine if a face has left the camera?  How do you handle those frames in which faces are temporarily not detected?  And no doubt more besides...

I don't doubt that this is a frequent issue though, so I'm sure people have come up with solutions before...  So it's either a matter of finding them; or hoping they post something here; or figuring something out yourself...

Best of luck!